Effects of tethered ligands and of metal oxidation state on the interactions of cobalt complexes with the 26S proteasome

Article abstract of DOI:10.1016/j.jinorgbio.2011.09.013

In this paper we report on the synthesis and characterization of three cobalt complexes described as [Co^II(L¹)₂] (1), [Co^II(L²)] (2), and [Co^III(L¹) ₂]ClO₄ (3). These complexes contain the deprotonated forms of the [NN′O] tridentate ligand HL¹ and its newly synthesized [N₂N′₂O₂] hexadentate counterpart H ₂L², namely, 2,4-diiodo-6-((pyridine-2-ylmethylamino) methyl)phenol and 6,6′-((ethane-1,2-diylbis((pyridin-2-ylmethyl) azanediyl))bis(methylene))bis(2,4-diiodophenol). Characterizations for 1-3 include electrospray ionization (ESI) spectrometry, infrared, and UV-visible spectroscopies, and elemental analyses. A detailed ¹H-NMR study was conducted for 3 and X-ray structural data was obtained for 2. The viability of this series as potential agents for proteasome inhibition and cell apoptotic induction involving PC-3 cancer cells is presented comparing the behavior of the untethered [NN′O]₂ six-coordinate 1 and 3 and the tethered counterpart 2 with a 1:1 metal-to-ligand ratio. It is observed that the tethering in 2 decreases inhibition activity. When 1 and 3 are compared, the most inert, but redox-active, cobalt(III) species shows the highest chymotrypsin-like activity inhibition on purified proteasome and PC-3 cancer cells. A hypothesis based on the role of oxidation states for proteasome inhibition is offered.

Full text of DOI:10.1016/j.jinorgbio.2011.09.013

Journal of Inorganic Biochemistry 105 (2011) 1759–1766
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Effects of tethered ligands and of metal oxidation state on the interactions of cobalt
complexes with the 26S proteasome
Dajena Tomco a, Sara Schmitt b, Bashar Ksebati a, Mary Jane Heeg a, Q. Ping Dou b, Cláudio N. Verani a,
⁎
a
Department of Chemistry, Wayne State University, 5101 Cass Ave. Detroit, MI 48202, USA
b
The Prevention Program, Ann Barbara Karmanos Cancer Institute, and Department of Pathology, School of Medicine, Wayne State University, Detroit, MI 4820, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 June 2011
Received in revised form 7 September 2011
Accepted 8 September 2011
Available online 16 September 2011
In this paper we report on the synthesis and characterization of three cobalt complexes described as
[CoII(L1)2] (1), [CoII(L2)] (2), and [CoIII(L1)2]ClO4 (3). These complexes contain the deprotonated forms of
the [NN′O] tridentate ligand HL1 and its newly synthesized [N2N′2O2] hexadentate counterpart H2L2, namely,
2,4-diiodo-6-((pyridine-2-ylmethylamino)methyl)phenol and 6,6′-((ethane-1,2-diylbis((pyridin-2-ylmethyl)
azanediyl))bis(methylene))bis(2,4-diiodophenol). Characterizations for 1–3 include electrospray ionization
(ESI) spectrometry, infrared, and UV–visible spectroscopies, and elemental analyses. A detailed 1H-NMR study
was conducted for 3 and X-ray structural data was obtained for 2. The viability of this series as potential agents
for proteasome inhibition and cell apoptotic induction involving PC-3 cancer cells is presented comparing the be-
havior of the untethered [NN′O]2 six-coordinate 1 and 3 and the tethered counterpart 2 with a 1:1 metal-to-ligand
ratio. It is observed that the tethering in 2 decreases inhibition activity. When 1 and 3 are compared, the most
inert, but redox-active, cobalt(III) species shows the highest chymotrypsin-like activity inhibition on purified
proteasome and PC-3 cancer cells. A hypothesis based on the role of oxidation states for proteasome inhibition
is offered.
Keywords:
Proteasome inhibition
Cobalt complexes
NMR spectroscopy
Cell apoptosis
© 2011 Elsevier Inc. All rights reserved.
1. Introduction
channels of the proteasome, or to the terminal threonines of the β ac-
tive sites in a similar way as reported to the antineoplastic agent bor-
Our groups have been interested in the development of coordina-
tion complexes capable of inhibiting the activity of the 26S protea-
some in tumorous prostate cells. The 26S proteasome is a large
protein complex responsible for the destruction of faulty proteins
and enzymes [1,2]. In tumor cells the activity of the 26S proteasome
goes in overdrive fostering the proteolysis of inhibition factors such
as IκB releasing NFκB nuclear factor that support the development
of blood vessels and promote tumor cell growth [3,4]. Inhibition of
this activity leads to cellular apoptosis, or programmed cell death.
We have demonstrated recently that a series of gallium complexes
[GaIII(LNN′O)2]+ (with the ligand 2,4-di-X-6-((pyridine-2-ylmethyla-
mino)methyl)phenol, where X=bromo or iodo) [5,6] were active
against the proteasome, as measured by the inhibition of its chymo-
trypsin-like activity (CT) and resulting accumulation of ubiquitinated
proteins. This activity triggered cell death both in vitro and in vivo.
Thus, we proposed that in order to inhibit the proteasome activity,
this species would likely bind to available amino acids along the α
tezomib (valcade) [7]. These [ML2]+ species would need to be
converted into a [ML]2+ species, in order to establish a metal/amino
acid chemical bond. Evidence for such mechanism was gathered
when related 2:1 and 1:1 copper(II) complexes showed comparable
CT inhibition activities in purified 20S and 26S proteasomes, in C4-
2B and PC-3 cell extracts, as well as in intact cells [8]. Consistently
we have also observed that [ML2] complexes with labile zinc(II)
ions show considerable activity, whereas more inert and redox-inac-
tive metal ions such as nickel(II) show negligible results [9].
In this paper we further this investigation by comparing the inhi-
bition activity of three cobalt complexes described as [CoII(L1)2] (1),
[CoII(L2)] (2), and [CoIII(L1)2ClO4] (3) (Scheme 1).
We are interested in the comparison of 1 and 2 that allows further
analysis for the need of ligand dissociation, where the newly designed
ligand used in 2 hinders such dissociation. Similarly we want to com-
pare 1 and 3 to examine the role of the oxidation states and the po-
tential of redox changes in proteasome inhibition. The antitumor
activity of cobalt species is fairly understudied, [10] but activity
against prostate cancer has been recently demonstrated by McNail
et al. for bivalent cobalt. [11] On the other hand, Teicher, Ware,
Hambley, [12] and more recently Donelly [13] have taken advantage
of bioreductive activation of its trivalent counterpart for the treat-
ment of hypoxic tumors. In this case, intracellular redox conversion
of Co(III) into Co(II) releases alkylating mustards.
⁎
Corresponding author at: Department of Chemistry, Wayne State University, 501
Cass Ave, Detroit, MI 48202, USA. Tel.: +1 313 577 1076; fax: +1 313 577 8022.
E-mail address: cnverani@chem.wayne.edu (C.N. Verani).
0162-0134/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2011.09.013

1760
D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
Table 1
Crystal data and structure refinement results for [CoIIL2] (2).
[CoII(L2)] (2)
Formula
M
C28H24CoI4N4O2
1015.04
Space group
a / Å
b/ Å
C2/c
13.8089(4)
9.9958(4)
22.0818(9)
c/ Å
o
α/
β/
γ/
o
96.013(2)
o
V/ Å3
3031.20(19)
4
Z
T/ K
λ/ Å
100(2)
0.71073
2.224
4.673
3.13
Scheme 1. Cobalt complexes.
Dcalc/ g cm-3
μ/ mm−1
R(F) (%)
Rw(F) (%)
2. Experimental
6.82
2.1. Materials and methods
aR(F)=∑(Fo!-!Fc( ∕ ∑!Fo! for IN2 s(I); Rw(F)=[∑w(Fo2 – Fc2)2 / ∑w(Fo2)2]1/2 for
IN2 s(I).
All of the reagents were used as received from commercial
sources. Methanol was dried using calcium hydride, and dichloro-
methane was doubly purified using alumina columns in an Innovative
Technologies solvent purification system. Infrared spectra were mea-
sured from 4000 to 400 cm−1 as KBr pellets on a Bruker Tensor 27
FTIR spectrophotometer. Electrospray ionization mass spectra in the
positive mode (ESI positive) were measured in a Micromass Quattro
LC triple quadrupole mass spectrometer, and experimental assign-
ments were simulated for peak position and isotopic distribution. El-
emental analyses were performed by Midwest Microlab: Indianapolis,
Indiana. Visible spectroscopy from 1.0×10−4 M N,N-dimethylforma-
mide solutions were performed using a Cary 50 spectrometer in the
range of 250 to 1100 nm. Trypan blue exclusion dye was purchased
from Sigma Aldrich (St. Louis, MO). The peptide substrate SucLL-
VYAMC (for the proteasomal chymotrypsinlike activity) was pur-
chased from Calbiochem, Inc. (San Diego, CA). RPMI 1640, penicillin,
and streptomycin were purchased from Invitrogen (Carlsbad, CA).
Fetal bovine serum was purchased from Aleken Biologicals (Nash,
TX). Antibodies against ubiquitin, actin, and secondary antibodies
were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Mouse monoclonal antibody against human PARP was purchased
from BIOMOL International LP (Plymouth Meeting, PA).
2.4. Biological assays
2.4.1. Proteasomal activity in purified 20S proteasome
Purified human 20S proteasome (35 ng) was incubated with
10 μM of CT substrate, SucLLVYAMC, in 100 μL of assay buffer
[20 mM Tris–HCl (pH 7.5)] in the presence of complexes 1–3 at vari-
ous concentrations, as well as the salt [Co(H2O)6](ClO4)2 and DMSO
solvent as control experiments. After 2 h of incubation at 37 °C, the
production of hydrolyzed AMC groups was measured using a Wallac
Victor3 multilabel counter with an excitation filter of 365 nm and
an emission filter of 460 nm [16].
2.4.2. Cell cultures and whole cell extract preparation
PC-3 human prostate cancer cells were grown in RPMI1640 sup-
plemented with 10% fetal bovine serum and maintained at 37 °C
and 5% CO2. A whole cell lysate was prepared as previously described
[17].
2.4.3. Cell proliferation assay
Cells were seeded in quadruplicate in a 96 well plate and grown to
70−80% confluence, followed by treatment with the indicated agents
for 18 h followed by measurement of cell proliferation by the 3
(4,5dimethylthiazol-2-yl)2,5diphenyltetrazolium bromide (MTT)
assay as described previously [18].
2.2. X-ray structural determination of [CoIIL2]
Diffraction data were measured on a Bruker X8 APEX-II kappa ge-
ometry diffractometer with Mo radiation and a graphite monochro-
mator. Frames were collected at 100 K with the detector at 40 mm
and 0.3° between each frame and were recorded for 10 s. APEX-II
[14] and SHELX [15] software were used in the collection and refine-
ment of the model. Crystals of [CoIIL2] (2) appeared as amber plates.
A total of 23770 reflections were measured, yielding 3750 unique
data (Rint=0.049). Hydrogen atoms were placed in calculated posi-
tions. The asymmetric unit contains one neutral complex without sol-
vate. The complex crystallized in 1:1 ratio of dichloromethane and
methanol solution (Table 1).
2.4.4. Proteasome CT-like activity in cells
Proteins extracted from cells after each treatment were incubated
for 2 h at 37 °C in 100 μL of assay buffer (50 mM Tris–HCl, pH 7.5)
with 10 μM of fluorogenic CT substrate SucLLVYAMC, as described
previously [17].
2.4.5. Western blot analysis
Cell extracts were separated by SDSPAGE and transferred to a ni-
trocellulose membrane. A western blot analysis was performed
using specific antibodies to PARP or ubiquitin, followed by visualiza-
tion using the HyGLO reagent (Denville Scientific, Metuchin, NJ).
2.3. NMR spectroscopic measurements
NMR spectra were measured using a Varian Mercury-400 and
VNMRS-500 MHz spectrometers. Proton, carbon-13, distortionless
enhancement by polarization transfer (DEPT), homonuclear correla-
tion spectroscopy (COSY), and heteronuclear multiple quantum co-
herence (HMQC), and nuclear Overhauser effect (NOE) spectroscopy
experiments were acquired in CDCl3, DMSO-d6 at 298 K and DMF-d7
between 233 and 273 K.
2.5. Syntheses
2.5.1. Ligand syntheses
The iodo-substituted ligand HL1 was synthesized as previously de-
scribed [19,20] by the treatment of 2-hydroxy-3,5-diiodobenzaldehyde
with pyridin-2-ylmethanamine in methanol followed by reduction
with sodium borohydride.

D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
1761
The organic precursor 2,4-di-iodo-6-(chloromethyl)phenol was
observed. The solution gave orange colored crystals after 48 h.
Yield: 65%; IR data: (KBr, cm−1, s = strong, m = medium) 2839
(m) (C-H), 1558(s), 1483(s) (C=NPy, C=CAr); 1363(m) (C-O);
ESI+ in MeOH: m/z (100%)=1015.7 for [CoIIL2]+; Anal. Calc. for 2
(%, C28H24CoI4N4O2, FW=1014.74 g.mol−1); C, 33.13; H, 2.38; N,
5.52. Found (%): C, 33.25; H, 2.44; N, 5.44.
synthesized as previously reported [21]. Yield: 90%. IR (KBr, cm−1
,
s=strong) 1450(s) (C=Car); 1265 (s) (C−O); 1H−NMR [δ, ppm;
s=singlet, 400 MHz, CDCl3, 300 K]=4.60, [2x s 2x 1H (CH2)]; 7.60
[s, 1 H (aryl) ]; 7.93 [s, 1 H (aryl)].
The new tethered ligand 6,6′-((ethane-1,2-diylbis((pyridin-2-
ylmethyl)azanediyl)) bis(methylene))bis(2,4-diiodophenol) H2L2
was synthesized by adapting available procedures for similar species
[22]. The precursor N1,N2-bis(pyridin-2-ylmethyl)ethane-1,2-di-
amine was obtained when a 40 mL methanolic solution (0.9 g,
15 mmol) of ethane-1,2-diamine, was treated with two equivalents
of picolinaldehyde (3.21 g, 30 mmol) and refluxed for 2 hours. The
resulting yellow Schiff base was reduced by addition of sodium boro-
hydride (1.4 g, 37 mmol) at 0 °C. The solvent was rotoevaporated and
the crude product was dissolved in 100 mL of brine. Extraction with
4×25 mL of dichloromethane followed, and the combined extracts
were dried over MgSO4. The solution was rotoevaporated and the
product was obtained as a viscous oil.
[CoIII(L1)2]ClO4 (3). The [Co(H2O)6](ClO4)2 salt (0.37 g, 1.0 mmol)
was dissolved in 5 mL of methanol and added to a 30 mL dichloro-
methane solution containing the ligand HL1 (0.98 g, 2.0 mmol) and
Et3N (0.28 mL; 2.0 mmol) under aerobic condition. The solution was
refluxed for 4 h, when a brownish product was vacuum filtered and
washed with cold methanol and ether. Yield: 70%; IR data (KBr,
cm−1, s = strong, m = medium): 1603(m), 1560(s), 1446(s)
(C=NPy, C=CAr); 1099(s) cm−1 (ClO4-) 1328(m) (C-O); ESI+ in
MeOH: m/z (100%)=989 for [CoIII(L1)2]+ ; Anal. Calc. for 3 (%,C26H22-
CoI4N4O6Cl, FW=1088.48 g.mol−1) C, 28.69; H, 2.04; N, 5.15. Found
(%): C, 28.72; H, 2.17; N, 5.05.
The precursor N1,N2-bis(pyridin-2-ylmethyl)ethane-1,2-diamine
(0.31 g, 1.3 mmol) was dissolved in 40 mL dichloromethane, when
triethylamine base (0.6 g, 6.4 mmol) was added dropwise and fol-
lowed by the addition of 2-(chloromethyl)-4,6-diiodophenol (1.0 g,
2.5 mmol). The solution was refluxed for 48 h. The product was
extracted with 3×25 mL of dichloromethane to remove the triethy-
lammonium chloride salt. The organic layer was dried over MgSO4
and rotoevaporated to give a yellow amorphous foamy solid which
was then recrystallized in methanol to yield a microcrystalline solid.
Yield: 75%. Mp 158–160 °C; IR (KBr, cm−1, s=strong, m=medium)
2818(s) (C-H), 1596(s), 1540(s) (C=NPy, C=CAr); 1363(m) (C-O);
1250 (s), (C-N); 1H NMR [δ, ppm; s=singlet, d=doublet, dd=doub-
let of doublet, 400 MHz, CDCl3, 300 K] 2.67, [4x s, 4x 1H (2 CH2) -N-
CH2-CH2-N-]; 3.61, [4x s, 4x 1H (2 CH2) –N-CH2-phenol]; 3.72, [4x s,
4x 1H (2 CH2) –N-CH2-pyridyl]; 7.11-7.13, [4x 4x 1H (phenol)];
7.20-7.23, [2x dd, 2x 1H (py-2)]; 7.66-7.70, [2x dd, 2x 1H (py-3)];
7.91, [2x d, 2x 1H (py-4)]; 8.59, [2x d, 2x 1H (py-1)]; ESI+ in
3. Results and discussion
3.1. Synthesis and characterization
Both ligands HL1 and H2L2 were synthesized in moderate yields
following procedures available in the literature [21,22]. The conden-
sation reaction of picolinaldehyde with ethane-1,2-diamine produced
the corresponding Schiff base ligand. This was followed by reduction
with sodium borohydride to generate the intended N1,N2-bis(pyri-
din-2-ylmethyl)ethane-1,2-diamine precursor. Formation of the
new H2L2 ligand, as shown in Scheme 2, was obtained by reaction
of this amino-pyridine precursor with 2-(chloromethyl)-4,6-diiodo-
phenol in the presence of base.
The 1H-NMR spectra of the ligands were measured and will be rel-
evant for future comparisons. The spectrum for HL1 was recorded in
DMSO-d6 in order to compare with the 1H NMR of 3. A complete char-
acterization of HL1 ligand including the 1H-NMR spectra and 1H-1H
COSY is given in the Supporting Materials (Figures S1, S2, S3). From
the 1H-NMR spectra, the two methylene protons from both the pyri-
dine rings (H5, H5′) and phenol (H6, H6′) appear as two distinct sin-
glet resonances centered at 3.83 ppm and 3.87 ppm, respectively,
whereas the aromatic protons lie between the 7.28–8.60 ppm region.
The exchangeable protons from the aliphatic nitrogen atom (H9) and
the phenol group (H10) are observed as a broad resonance between
6.60 and 7.10 ppm. The (H1) proton from the pyridine ring resonates
as a doublet at 8.54 ppm.
MeOH: m/z (100%)=959.1 for [H2L2 +H+ +
] .
2.5.2. Complex syntheses
Note: Perchlorate salts of the metal complexes are potentially explo-
sive and should be handled with care.
[CoII(L1)2] (1). Synthesis of this complex followed the previously
described procedure [21]. Yield: 85%; IR data: (KBr, cm−1, s=strong,
m=medium) 1603(m), 1560(s), 1446(s) (C=NPy, C=CAr); 1328
(m) (C-O); ESI+ in MeOH: m/z (100%)=990.1 for [CoII(L1)2 +
+
H+
]
;
Anal. Calc. (%) for 1·H2O (%, C26H24CoI4N4O3,
The spectrum for H2L2 recorded in CDCl3, provides evidence for a
symmetric ligand in the solution, as Orvig et al. [23] and Neves et al.
[24] have similarly observed with other related H2bbpen ligands.
Therefore, three sharp singlet resonances are shown at 2.67 ppm,
3.61 ppm, and 3.72 ppm representing the four protons from ethane-
1,2-diamine, phenolic-methyl, and pyridyl-methyl groups, respectively;
whereas the aromatic protons are displayed between the region of 7.11–
8.60 ppm.
FW=1007.04 g.mol-1) C, 31.01; H, 2.40; N, 5.56. Found (%): C,
31.14; H, 2.16; N, 5.39.
[CoIIL2] (2). [Co(H2O)6](ClO4)2 (0.12 g, 0.31 mmol) was dissolved
in 5 mL of methanol and was added dropwise under anaerobic condi-
tions to
a 10 mL dichloromethane solution containing (0.3 g,
0.31 mmol) of the H2L2 ligand and Et3N (0.079 g; 79.2 mmol). The so-
lution was stirred at room temperature for 3 h, when a color change
was observed from yellow to brown. No precipitate was formed
Scheme 2. Synthesis of the H2L2 ligand.

1762
D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
Table 2
1H NMR assignment for HL1 and [Co(L1)2]ClO4 (3) in DMSO-d6.
Proton H
HL1 (ppm)
[CoIII(L1)2ClO4] (ppm)
H1
H2
H3
H4
8.54 (d)
8.99 (dd)
7.46 (t)
7.95 (t)
7.29 (dd)
7.8 br (dt)
7.37 (d)
7.57 (d)
H5 H5′
H6 H6′
H7
3.83 (s)
3.87 (s)
7.33 (d)
4.39(d) 5.58 (dd)
3.48(dd) 3.66 (d)
7.39 (s)
H8
7.78 (d)
7.34 (s)
H9
H10
6.60–7.10 br (s)
6.60–7.10 br (s)
8.33 (s)
No peak
Spectrometric analysis for complexes 1–3 was performed by ESI-
MS in the positive mode and in methanol solutions. The results
showed excellent agreement between the experimental and the sim-
ulated data as presented in Fig. 1. As previously published, [21] com-
plex 1 shows the presence of the molecular ion peak observed at
+
m/z+ =989.9 (100%) which corresponds to the {[Co(L1)2]+H+
}
species. Alternatively, for complex 3 the main peak is detected at
m/z+ =988.9 (100%) and assigned to the [Co(L1)2]+ species provid-
ing strong evidence for the coordination of the cobalt(III) ion. ESI-
MS for complex 2 shows the parent peak for {[Co(L2)]+H+ + species
}
at m/z+=1015.7 (100%). The isotopic distribution for both complexes
2 and 3 are shown in Fig. 1. All of the above data confirm the expected
configurations of 1:2 and 1:1 metal-to-ligand species containing six-co-
ordinate cobalt ions bound to HL1 or H2L2, respectively.
Fig. 1. Isotopic distribution for complex 2 (left) and 3 (right). The bars indicate the ex-
perimental results and the continuous spectra indicate the simulated results.
Complexes 1–3 were synthesized by the reaction of the metal salt
with the appropriate ligand in methanol or dichloromethane in the
presence of triethylamine base. Characterization of all of the metal
complexes includes spectroscopic and spectrometric techniques con-
sisting of IR, NMR, ESI-MS, and elemental analyses. Infrared analysis
confirms that both C=N and C=CAr stretching modes shift ca. 40–
60 cm−1 to lower frequencies due to metal coordination. Characteristic
for the infrared spectrum of the [CoIII(L1)2]ClO4 (3) complex is the pres-
ence of a strong broad band at 1098 cm−1, corresponding to the per-
chlorate counterion.
3.2. Molecular structural characterization of [CoIIL2] (2)
X-ray diffraction was used to determine the molecular structure of
single crystals of compound 2 isolated from slow solvent evaporation
of a 1:1 mixture of dichloromethane and methanol. Fig. 2 displays the
ORTEP diagram for complex 2 with selected bond lengths and angles.
Compound 2 consists of a neutral cobalt(II) species with coordination
environment arranged about a fully deprotonated hexadentate ligand
(L2)2-, containing two [NN′O] donor sets. The pseudo-octahedral
Fig. 2. ORTEP diagram at 50% probability level for 2. Selected bond lengths include Co(1)-O(1)=2.003(3), Co(1)-N(1)=2.139(3), Co(1)-N(2)=2.234(3) Å. Selected angles include
N(1)-Co(1)-N(2)=76.57(12), O(1)-Co(1)-N(1)=90.32(12), O(1)-Co(1)-N(2)=89.65(11). Goodness of fit is given by R(F) (%)=3.13.

D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
1763
b
DMSO-d6
H2O
a
8
7
3
4
2
1
5
6’
9
6
5’
9
8
7
6
5
4
3
PPM
Fig. 3. NMR spectroscopic measurements for 3; (a) 1H-NMR spectrum and (b) HMQC spectra.
cobalt(II) center is oriented in a two-fold rotation axis with each half
of the ligand in a facial arrangement. We have previously reported on
similar coordination spheres for cobalt(II) complexes containing two
independent [NN′O] ligands, in which all equivalent donors are posi-
tioned trans to each other [21,25]. In contrast, complex 2 displays an
arrangement of cis phenolates, trans pyridines, and cis nitrogen
atoms, rather observed for iron(III) and manganese(III) complexes
with similar hexadentate ligand [26,27,28]. The short cobalt-donor
bonds, along with the absence of counterions support the bivalent na-
ture of the metal ion. The observed Co-Ophenolate distance is 2.00 Å,
whereas Co-Npyridine is 2.14 Å, and Co-Namine is 2.23 Å. These values
are in good agreement with the literature [21].
3.3. 1H NMR spectroscopic studies
The 1H-NMR data for ligands HL1 and H2L2 has been discussed in
Section 3.1. The spectra for compound 3 were taken in DMSO-d6 at
298 K, and are shown in Fig. 3. A comparison of the proton chemical
shifts of HL1 and 3 is given in Table 2 and (Table S1). The 1H-NMR
spectrum of 3 showed eleven distinctive resonances between 3.41
and 9.00 ppm, indicating that the compound has a diamagnetic
3d6
electronic configuration. The 13C-NMR spectrum of com-
low-spin
plex 3 confirmed the presence of thirteen different carbon chemical
Control
# 1
140
120
100
80
# 3
60
40
20
0
DMSO
10 µM
20 µM
30 µM
40 µM
50 µM
Fig. 5. MTT, PC-3 cells after 18 h treatment. Control is DMSO, for each concentration
from 10 to 50 μM, compound 1 is indicated in the left column and 3 is indicated in
the right column.
Fig. 4. UV–visible spectra of complexes 1–3 in N,N-dimethylformamide, 1.0×10−4 M.

1764
D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
Fig. 6. Chymotrypsin-activity inhibition in human purified 20S proteasome; DMSO and CoII(ClO4)2 are controls. Top: comparison between 1 (left column) and 2 (right column);
bottom comparison between 1 (left column) and 3 (right column).
peaks corresponding with half the number of carbons present in the
molecule, suggesting that this complex is symmetric in solution. A
combination of DEPT, COSY (Figures S4 and S5), and HMQC experi-
ments were used to establish proton and carbon connectivity in the
isolated spin systems for 3, whereas NOE was employed to confirm
the proposed structure. DEPT analysis revealed two signals corre-
sponding to two methylene groups, and six distinct CH peaks giving
a total of ten protons, whereas 13C-NMR/DEPT studies confirmed
the presence of five fully-substituted aromatic carbons. Characteristic
for the 1H NMR spectra for complex 3 is the observation of the chem-
ical shifts and splittings of the proton resonances attributed to metal
coordination for the two methylene protons (Fig. 3a). The original
singlet peak observed at 3.83 ppm (H5) in the non-metallated ligand
now appears as the two doublets observed at 4.39 and 5.58 ppm for
the complex, possibly due to the coupling with the aliphatic amine
proton (H9). The same phenomenon was observed for the initial sin-
glet peak at 3.87 ppm (H6), which transforms into two distinct reso-
nances appearing as doublets at 3.48 and 3.66 ppm for 3. The
connectivity of the protons to the carbons for complex 3 was sup-
ported by the HMQC experiment, as shown in Fig. 3b and Figure S6;
where the identity of the exchangeable proton attached to the ali-
phatic nitrogen atom (H9) was revealed. In this spectrum, one reso-
nance corresponding to the proton on the aliphatic amine (H9)
appears as a broad singlet and is shifted downfield to 8.33 ppm.
Also, in comparison to the 1H NMR spectra of the ligand HL1, where
the exchangeable proton of the phenol group is centered between
6.60 and 7.10 ppm, complex 3 has no such peak present which indi-
cates ligand deprotonation. The aromatic proton found on the 1-posi-
tion of the pyridine ring (H1) shifts downfield to 8.99 ppm and
appears as a doublet of a doublet due to metal coordination. Only
slight changes of the proton chemical shifts were observed for all ar-
omatic protons as shown in Table 2, and no broadening of the signal
indicative of dynamic equilibria [29] were observed. To confirm the
thermodynamic stability of 3 in solution, variable-temperature 1H-
NMR experiments were performed in DMSO-d6 (25 to 80 °C, not
shown), as well as in DMF-d7 (25 to −40 °C, Figure S7). No significant
changes were observed, thus supporting the idea that the 3d6
low spin
assignment for 3 is the stable conformation in the timeframe of
these experiments.
3.4. UV–visible spectroscopy
The UV–visible spectra of complexes 1–3 were taken in 1.0×10−4 M
DMF solutions to assure consistency of solubility properties (Fig. 4). The
noticeable absence [21] of a ligand-to-metal charge transfer band at
about 450 nm is diagnostic of the 3d7high spin cobalt(II) center in com-
plexes 1 and 2. This very characteristic process is observed for complex
⁎
3 at 441 nm corresponding to a pπphenolate – dσ cobalt (III) charge trans-
fer. Another band at 630 nm usually attributed to d-d transitions is
also present, confirming the trivalent character of the 3d6
co-
⁎
low spin
balt(III) species. Other observed features include the intense π – π
intraligand bands at about 320 nm for all species. These results are in
good agreement with the previously published data on cobalt com-
plexes containing [NN′O] donor ligands [10,19].
3.5. Biological results
Biological data were gathered regarding the inhibition activity of
species 1–3 toward cell proliferation, the purified 20S proteasome, in-
tact PC-3 prostate cancer cells, and finally, toward tumor cell selectiv-
ity. Inhibition of cell proliferation in human PC3 cells was successful
with species 3. Cell proliferation was decreased by 40% upon treat-
ment with 20 μM, and reached nearly 100% inhibition at a concentra-
tion of 30 μM. A less remarkable activity was observed for 1, where
noticeable inhibition required considerably higher concentrations,
e.g. ~30% inhibition at 50 μM (Fig. 5). Initially, this observation
seems to contradict the proposed ligand exchange mechanism

D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
1765
#1
#3
120
100
80
60
40
20
0
Control
#1
NT DM 40 50 10 20 30 40 50 µM
#3
Ub-proteins
DMSO
10 µM
20 µM
30 µM
40 µM
50 µM
Fig. 7. Chymotrypsin-activity inhibition in PC-3 cell lysates after 18 h treatment. Com-
pound 1 was measured at 40 and 50 μM and is shown in the left column. Compound 3
is shown as a single column for 1–30 μM and in the right column for 40 and 50 μM.
DMSO is the control.
observed for similar copper, nickel and zinc complexes with similar li-
gands [8,9].
PARP
Actin
To test the in vitro 20S proteasome inhibition ability of these species, a
comparison of the inhibitory activity of 1, 2 and 3 to the 20S proteasomal
activity was performed under cell-free conditions. Human purified 20S
proteasome was incubated with the Co(ClO4)2·6H2O salt as control, as
well as with 1, 2, and 3 at different concentrations, followed by measure-
ment of CT-like activity. This activity was marginally inhibited by the li-
gands and was inhibited by the highest concentration (50 μM) of cobalt
salt at a 50% level. This result confirms that in 20S proteasome – where
the regulatory 19S caps have been removed – metal ions will show
some inhibitory effect, as previously observed for gallium salts [30]. In
spite of modest activities, comparison between complexes 1 and 2, dis-
played in Fig. 6 (top) shows the highest inhibition for 1 (40% at 10 μM),
where two independent ligands are present. This result reinforces the
notion that facilitated ligand exchange will foster inhibition. No activity
is observed for 2 even at 50 μM.
Fig. 9. Western blot for PC-3 cell lysates after 18 h treatment.
associated with apoptotic cell death, apoptotic-specific caspase-3 in-
duction (Fig. 8) and related PARP disappearance (Fig. 9) were mea-
sured spectrophotometrically and by Western blotting, respectively.
Dramatic induction of caspase-3 was observed in cells treated with
3 at 40 μM. As expected, abrogation of full length PARP only occurred
in cells treated with 30–50 μM of 3, whereas cells treated with 1 at
the highest concentration tested had little visible effect. These results
show that induction of apoptosis by 3 in PC-3 cells is associated with
inhibition of proteasomal chymotrypsin-like activity.
The activities of 1 and 3 were also compared, as shown in Fig. 6
(bottom). These species show constant 1:2 metal-to-ligand ratio but
allow for the respective comparison between a labile 3 d7 and an
inert 3 d6 metal center. The latter is expected to display little or no in-
hibitory effects, as observed for the equally inert nickel(II) ion [9]. Un-
expectedly, species 3 displays remarkably better CT-like activity
inhibition of the 20S core than its labile counterpart.
Intrigued by the previous results, proteasome inhibition and apo-
ptosis induction were tested in intact PC-3 human prostate cancer
cells to confirm the potential activity of 3. Cells were first treated
with different concentrations (up to 50 μM) of 1 and 3 for 18 h, fol-
lowed by measurement of proteasome inhibition. PC-3 cells treated
with 3 showed a dose-dependent inhibition of the proteasomal activ-
ity by ~35% at 30 μM and ~95% at 50 μM (Fig. 7). Consistently, levels
of ubiquitinated proteins were increased in a dose-dependent man-
ner in PC-3 cells for 3, whereas 1 showed negligible inhibition.
Previous reports have indicated that inhibition of proteasomal
chymotrypsin-like activity in tumor cells may result in the induction
of apoptosis [31]. To investigate whether proteasome inhibition is
These results confirm the need for ligand dissociation, as shown by
comparison between the activities of 1 and 2, both with a labile 3 d7
configuration. However, it is puzzling that compound 1 has proven
to be less active than the inert 3 d6 metal-containing complex 3. Un-
like the previously studied inert 3 d8 nickel(II) ion, cobalt(III) is a
redox-active species capable of being reduced to cobalt(II) within the
reducing cellular environment by available reductants. This reduction
has been demonstrated individually by several groups [12,13] and uti-
lized for release of alkylating agents such as nitrogen mustards. The
redox potential for the cytosolic environment is reported to be around
−0.3 V vs. NHE by Østegaard [32]. A preliminary cyclic voltammo-
gram for 3 in DMSO/TBAPF6 suggests the Co(III)/Co(II) couple at
around −0.5 V vs. NHE. Although more negative, the result suggests
bioreductive activation as a valid working hypothesis in need of fur-
ther exploration.
4. Conclusions
In this paper we have investigated the interaction of cobalt com-
plexes with the 26S proteasome. We have compared the behavior of
1:2 and 1:1 metal-to-ligand six-coordinate cobalt species toward
cell proliferation, the purified 20S proteasome, and intact PC-3 pros-
tate cancer cells. The 1:2 species described as 1 and 3 are formed re-
spectively between cobalt(II) or cobalt(III) ions and two
deprotonated [NN′O] ligands (L1)-, whereas the 1:1 species 2 is
based on a cobalt(II) ion and the new [N2N′2O2] ligand H2L2 in its
deprotonated form. Detailed characterization along with a crystal
structure allows for unquestionable identification of 2, whereas me-
ticulous NMR spectroscopic evaluation indicated configurational sta-
bility of 3 from 80 to −40 °C. The CT-like activity inhibition is
severely hampered for the 1:1 species 2, reinforcing the current hy-
pothesis of ligand dissociation as a requirement for proteasome
9
8
7
6
5
4
3
2
1
0
Control
#1
#3
DMSO
10 µM
20 µM
30 µM
40 µM
50 µM
Fig. 8. Comparison between 1 (left column) and 3 (right column) for Caspase-3 (apo-
ptosis) Induction in PC-3 cell lysates after 18 h treatment. DMSO is the control.

1766
D. Tomco et al. / Journal of Inorganic Biochemistry 105 (2011) 1759–1766
inhibition. Surprisingly, the kinetically inert 3 showed remarkable
proteasome inhibition, far superior to that observed for the labile 1.
We hypothesize that this difference is due to the fact that cobalt is
redox-active and species 3 is likely to be reduced intracellularly. In
this process a labile cobalt(II) species would be generated, favoring li-
gand dissociation and interaction with the proteasome. We also hy-
Appendix A. Supplementary data
Supplementary data to this article can be found online at doi:10.
1016/j.jinorgbio.2011.09.013.
References
pothesize that species 1, already containing
a labile cobalt(II)
species will not remain intact intracellularly in order to reach the tar-
geted proteasome. The possibility of using redox-active metals that
can be intracellularly bioreduced opens a stimulating window of op-
portunity to explore proteasome inhibition, both by metal activation
as demonstrated here and as suggested by Scarpellini [10] as well as
by using metal ions as carriers for drug delivery [12]. Such studies
are currently under investigation in our group.
[1] P. Kaiser, L. Huang, Genome Biology 6 (2005) 233.
[2] M. Hochstrasser, Current Opinion in Cell Biology 7 (1995) 215–223.
[3] S. Meiners, A. Ludwig, V. Stangl, K. Stangl, Medicinal Research Reviews 28 (2008)
309–327.
[4] L. Borissenko, M. Groll, Chemistry Review 107 (2007) 687–717.
[5] R. Shakya, F. Peng, J. Liu, M.J. Heeg, C.N. Verani, Inorganic Chemistry 45 (2006)
6263–6268.
[6] M. Frezza, C.N. Verani, D. Chen, Q.P. Dou, Letters in Drug Design and Discovery 4
(2007) 311–317.
[7] A.F. Kisselev, A.L. Goldberg, Chemical Biology 8 (2001) 739–758.
[8] S. Hindo, M. Frezza, D. Tomco, M.J. Heeg, L. Hryhorczuk, B.R. McGarvey, Q.P. Dou,
C.N. Verani, European Journal of Medicinal Chemistry 44 (2009) 4353–4361.
[9] M. Frezza, S.S. Hindo, D. Tomco, M. Allard, Q.C. Cui, M.J. Heeg, D. Chen, Q.P. Dou, C.N.
Verani, Inorganic Chemistry 48 (2009) 5928–5937.
[10] E.T. Souza, L.C. Castro, F.A.V. Castro, L. do Canto Visentin, C.B. Pinheiro, M.D. Pereira, S.
de Paula Machado, M. Scarpellini, Journal of Inorganic Biochemistry 103 (2009)
1355–1365.
Abbreviations
AMC
COSY
CT
7-amino-4-methylcoumarin
homonuclear correlation spectroscopy
chymotrypsin
DEPT
DMF
distortionless enhancement by polarization transfer
dimethylformamide
[11] L. Gurley, N. Beloukhina, K. Boudreau, A. Klegeris, W.S. McNeil, Journal of Inorgan-
ic Biochemistry 105 (2011) 858–866.
ESI-MS electron spray ionization mass spectrometry
[12] M.D. Hall, T.W. Failes, N. Yamamoto, T.W. Hambley, Dalton Transactions (2007)
3983–3990.
[13] C. Schieber, J. Howitt, U. Putz, J.M. White, C.L. Parish, P.S. Donnelly, S.-S. Tan, Jour-
nal of Biological Chemistry 286 (2011) 8555–8564.
[14] APEX II collection and processing programs are distributed by the manufacturer,
Bruker AXS Inc, Madison WI, USA, 2009.
[15] G.M. Sheldrick, Acta Crystallograhica A64 (2008) 112–122.
[16] K.G. Daniel, P. Gupta, R.H. Harbach, W.C. Guida, Q.P. Dou, Biochemical Pharmacol-
ogy 67 (2004) 1139–1151.
FTIR
Fourier-transform infrared spectroscopy
2,4-diiodo-6-((pyridine-2-ylmethylamino)methyl)phenol
6,6′-((ethane-1,2-diylbis((pyridin-2-ylmethyl)azanediyl))
bis(methylene))bis(2,4-diiodophenol)
heteronuclear multiple quantum coherence
3(4,5dimethylthiazol-2-yl)2,5diphenyltetrazolium
bromide
nuclear Overhauser effect
1640 (Roswell Park Memorial Institute) cell culture
medium
HL1
H2L2
HMQC
MTT
[17] D. Chen, Q.C. Cui, H. Yang, Q.P. Dou, Cancer Research 66 (2006) 10425–10433.
[18] K.G. Daniel, D. Chen, S. Orlu, Q.C. Cui, F.R. Miller, Q.P. Dou, Breast Cancer Research
7 (2005) R897–R908.
NOE
RPMI
[19] R. Shakya, C. Imbert, H.P. Hratchian, M. Lanznaster, M.J. Heeg, B.R. McGarvey, M.M.
Allard, H.B. Schlegel, C.N. Verani, Dalton Transactions (2006) 2517–2525.
[20] C. Imbert, H.P. Hratchian, M. Lanznaster, M.J. Heeg, L.M. Hryhorczuk, B.R. McGarvey,
H.B. Schlegel, C.N. Verani, Inorganic Chemistry 44 (2005) 7414–7422.
[21] R. Shakya, S.S. Hindo, L. Wu, M. Allard, M.J. Heeg, H.P. Hratchian, B.R. McGarvey, S.
da Rocha, C.N. Verani, Inorganic Chemistry 46 (2007) 9808–9818.
[22] A. dos Anjos, A.J. Bortoluzzi, B. Szpoganicz, M.S.B. Caro, G.R. Friedermann, A.S.
Mangrich, A. Neves, Inorganica Chimica Acta 358 (2005) 3106–3114.
[23] E. Wong, S. Liu, S.J. Rettig, C. Orvig, Inorganic Chemistry 34 (1995) 3057–3064.
[24] A. Neves, S.M.D. Erthal, I. Vencato, A.S. Ceccato, Y.P. Mascarenhas, O.R. Nascimento,
M. Hörner, A.A. Batista, Inorganic Chemistry 31 (1992) 4749–4755.
[25] D.L. Kepert, Inorganic Stereochemistry, Springer-Verlag, New York, 1982, p. 114.
[26] I.A. Setyawati, S.J. Rettig, C. Orvig, Canadian Journal of Chemistry 77 (1999)
2033–2038.
SDSPAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis
CDCl3
PARP
PC-3
NHE
deuterated chloroform
poly (ADP-ribose) polymerase
prostate cancer cell line
normal hydrogen electrode
TBAPF6 tetrabutylammonium hexafluorophosphate electrolite
Acknowledgments
[27] M. Lanznaster, A. Neves, A.J. Bortoluzzi, A.M.C. Assumpção, I. Vencato, S.P.
Machado, S.M. Drechsel, Inorganic Chemistry 45 (2006) 1005–1011.
[28] A. Neves, I. Vencato, S.M.D. Erthal, Inorganica Chimica Acta 262 (1997) 77–80.
[29] F. Silva, F. Marques, I.C. Santos, A. Paulo, A.S. Rodrigues, J. Rueff, I. Santos, Journal
of Inorganic Biochemistry 104 (2010) 523–532.
[30] D. Chen, M. Frezza, R. Shakya, C.Q. Cui, V. Milacic, C.N. Verani, Q.P. Dou, Cancer
Research 67 (2007) 9258–9265.
[31] B. An, R.H. Goldfarb, R. Siman, Q.P. Dou, Cell Death and Differentiation 5 (1998)
1062–1075.
The authors thankfully acknowledge support from the National
Science Foundation (Grants CHE-0718470 and CHE-1012413 for
C.N.V.), the Karmanos Cancer Institute Pilot Grant (KCI027354 for
C.N.V. and Q.P.D.), the Department of Defense Breast Cancer Research
Program (W81XWH-04-1-0688 and DAMD17-03-1-0175 for Q.P.D.)
and the National Cancer Institute (1R01CA120009 for Q.P.D.). D.T.
and S.S. respectively acknowledge graduate fellowships from the
WSU-College of Liberal Arts and Sciences and WSU-School of
Medicine.
[32] H. Østergaard, C. Tachibana, J.R. Winther, The Journal of Cell Biology 166 (2004)
337–345.